Product systems

Product systems are one of the more complex elements in openLCA. Main purpose of a product system is to model and calculate case studies.

Basically, a product system is a network of processes. The processes are connected to the reference process, which has the required product as output.

Creating a new product system

To build a new product system, open the wizard. Enter a name and specify a reference process. As an example we will create a case study of aluminium production.

If you check the marked box, the program will automatically build a whole product system, following up all the different connections in the database, starting from the reference process. Depending on the database, this function can lead to a very complex product system. If you need a small and clear system, you can build it manually; you can of course also modify a system that is created automatically, by adding and deleting processes manually.

If you click on „Finish”, the product system editor will be opened. At the bottom of the editor you find several tabs, for main elements of a product system

Elements in a product system

As you see from the screenshot above, each product system contains, beside general information, a graph, an LCI result, impact assessment results if available, and parameters.

Working with the outline

An element that is not in the tab band but belongs to a product system is the Outline. The outline lists all processes that are part of a product system. The outline window is opened by default when you open a product system; if you have closed it, you can re-open it via the menu by selecting Window/Show views/Outline (in English language of course).

The outline is convenient to locate processes in large product systems. You can filter and search for processes in the outline via entries in the top edit box of the outline...

..and you can show and hide selected processes, via the context menu (right mouse click):

openLCA tries to link a process to the existing visible process in the graphical editor. If this is not possible because connected processes are not visible at present, the to-be-shown process is placed below the other processes as separate box.

Working with the graphical editor

If you go to the tab Graph at the bottom of the product system editor, the product system will be shown as a network. If you start modeling a product system, only one process might be visible; if you have already connected several processes, the graph can look like this:

The graph editor is the central point to model product systems in openLCA.

Each process in the product system is represented as one box. You can move these boxes around by pressing the left mouse button and holding it while moving the object to the target location. This process is known also in other applications, as dragging. If you double-click on a process-box, the inputs and outputs are shown. Double-clicking again hides the inputs and outputs. You can also change the size of the process-boxes by dragging the frame. With the small minus and plus symbols in the processes you can expand or collapse the next „layer” of processes connected to the specific process.

Adding or deleting processes in the graph.
The graph is the visual structure of all processes that contribute to a product system. Adding or removing processes from this graph is an essential aspect of modelling the product system. If you add a new process it will be located top left, and you can then move it to another location and set a link from an output of the new process to an existing process that uses this output as input. If you want to delete a process in the graph you have to delete its links first. Just mark them and click the right mouse button. Then choose „Delete”. Now you can mark the process, click the right mouse button and say „Delete”.

If you decided to build the system automatically, you will initially see the process that provides the quantitative reference for the product system, plus the next layer of processes. If not, you will only see the reference process. By double-clicking on it, the process inputs and outputs are shown.

Now you can add further processes via drag and drop or via the search function. Note that every new process is placed at the upper left corner. If you add several processes successively, the boxes will lie on the top of each other.

In order to find the processes that either provide a product of a process, or receive products of the selected process, openLCA has an implemented search function. To use it, select a process-box and press the right mouse button, and you can search providers and recipients for all inputs and outputs of the selected process.

All appropriate processes are then displayed in a new list window. It is also indicated, on the right, whether the process is already part of the product system and whether it is already connected to the selected process (but not shown at present). If the process is not yet present in the product system, it can be added (check: ‚add to product system’), and it can also be linked to the selected process.

If you already know which process you want to connect, you can drag it in graph editor, from the process object window. Then click with the left mouse button on an input/output where you want to add a connection and hold it pressed while dragging it to the corresponding output/input of another process. All possible connections are highlighted in blue.

In these ways you can extend your product system as much as you want and as your databases contains appropriate processes.

The miniature view
A small miniature view window is available after a click on this icon in the menu bar

It always shows the compete graph of all processes, in a miniature view.

The window is useful for

• getting an overview of your processes
• navigating to a specific location of your process graph, by moving the grayed area in the window (which represents the part of the graph that is displayed in the large graph window) with your mouse
• changing the zoom factor of the graph window by moving the slider

Changing the layout
Processes can be moved around freely in the graphical editor. If you have expanded processes or minimised them, or added processes, it may be wise to let openLCA re-arrange the process boxes. Two layouts are available, one that creates a vertically oriented tree, and one that creates a more horizontal tree structure. The vertically oriented tree is called minimal tree. Both layouts are available via the layout icons,

Hints for working with very large product systems
Product systems, especially if they are automatically connected from large databases, can become very large. In these cases, the overview of the system structure may not be useful for modeling.

Instead, use the outline to look up processes, and to connect processes. Clicking on the minus icon in the menu will bring the expanded system to its origin state (the process that provides the quantitative reference and the next layer of connected processes). By adding processes via the outline, and connecting them as required, even extremely large product system with several thousand processes remain manageable in openLCA.

Calculating results

Inventory

To calculate the life cycle inventory or LCI results, go to the tab LCI results. You can select between two different calculation methods, matrix and sequential calculation. To do so, click the equation symbol at the top right at and select a method. Especially for large product systems with loops, the matrix method is recommended; the sequential method may be useful for Monte Carlo simulations with rather high truncation settings.

The sequential calculation method uses two different thresholds. These are important if the product system contains loops, i.e. if a process references itself (also indirectly, via other processes: heavy fuel oil that is used in a tanker that transports crude oil to a refinery that produces, among other things, heavy fuel oil).

These two thresholds are:

• the lower limit of the (additional) scaling factor; this limit considers the scaling factor calculated for the „next-to-consider” process, and does not follow this process if the threshold is below the specified value.
• the maximum number of iterations that are performed in a loop

Both these thresholds can be specified in the properties page for a product system. They are relevant only for the sequential calculation.

In the properties page you can also specify whether an uncertainty calculation shall be performed. The uncertainty calculation here is always performed by approximation formulas, using Gaussian first order error propagation. A Monte Carlo simulation is started with the simulation editor, described in a specific uncertainty simulation page (see section Advanced Functions).

These approximation formulas work best if the relative error is small. It has been found¹ that the approximation works well if the relative errors is below specific thresholds (for the calculation of process scaling factors, 20%) - this is checked during calculation, and if the relative error is higher, the result is presented with a warning.

Impact Assessment

In openLCA, impacts can be calculated only for product systems (and projects as comparisons of product systems). Processes are used within product systems, and not in itself available for impact calculation. It is however possible to create a product system with only one process and calculate the impacts then for this system.
If you want to analyse the impact assessment, first, the inventory of a product system needs to be calculated, as explained in the previous section. Save the inventory and then, in a next step, go to the next tab „LCIA results”. For obtaining the LCIA results, choose an impact assessment method and click on the calculation button.

If you click on you can export the results as an excel file.

The chart

Under the tab „Chart” you can picture the processes with the highest contributions to a certain impact category in a chart. To do so, select an LCIA method and an impact category you want to consider. Now enter a number x for the „top x” processes you want to picture and click on „Draw chart”.

The colours in the chart represent the process contribution: red = high impact, orange-yellow = medium impact, green = little impact, and white = no impact. If you click on one bar, the process name is displayed in a tooltip.

Further, you can visualise the top „x” flows with the highest impacts, in the same chart.

The Sankey diagram

openLCA draws a Sankey diagram of the product system. This diagram is a graphical illustration of the intensity of different flows in the product system.

Below the diagram in the section „Properties” you can edit the Sankey diagram. You can make flows visible or hide them (click on „Visible” and select „True” or „False”), add a limiting value for the process contribution you want to have displayed (click on „More than x percent” and enter a value), and change the flow colours (click on „Color”, then click on and select one).

In the section „Properties” you can find several symbols, useful for managing the display and properties of the Sankey diagram:

pins this property view to the current selection

shows the different categories

shows advanced properties

restores default value

hides unlinked processes

unselects all

changes the layout

Saving a product system as LCI result

A product system can be saved as an LCI result process by clicking the save as.. icon

You can select the process category where the process will be stored in the process category tree.

The so created process can be used in further product systems like any other process. Note that the new process represents the product system at one specific time (when the process was created); it does not reflect further modifications of the product system.

¹ Ciroth, A., Fleischer, G., Steinbach, J.: Uncertainty Calculation in Life Cycle Assessments – A Combined Model of Simulation and Approximation, Int J LCA 9 (4) 216 ? 226 (2004)